4016.p65 ALCES VOL. 40, 2004 BASKIN ET AL. - MOOSE ESCAPE BEHAVIOUR 123 MOOSE ESCAPE BEHAVIOUR IN AREAS OF HIGH HUNTING PRESSURE Leonid Baskin1, John P. Ball2,3, and Kjell Danell2 1Institute of Ecology and Evolution, Leninsky pr. 33, Moscow, 119071, Russia; 2Department of Animal Ecology, Swedish University of Agricultural Sciences, Umeå, SE-901 83, Sweden; 3e-mail: john.ball@szooek.slu.se ABSTRACT: Although hunters cause more than 80% of moose mortality in some geographic areas, quantitative studies of how moose attempt to escape humans are surprisingly rare. We experimen- tally disturbed radio-collared moose of known age and of both sexes to study escape behaviour from humans. We found that larger groups of moose made fewer stops between being disturbed and settling down, and that larger groups exhibited a longer path length before quieting. We detected no significant effect of age (a potential measure of survival rate) on escape behaviour. The escape path of males was significantly longer than females even though the linear distance from the site of disturbance to the location where the moose settled down was not significantly different between the sexes. Overall, the escape path of males from the site of disturbance to where they settled down was significantly more tortuous than that of females. Although males are the preferred prey of hunters, the differences in escape behaviour between the sexes also may contribute to why males are more frequently killed by hunters. Thus, in areas with heavy hunting pressure, hunters may be acting as a selective force that favours animals that immediately run away after disturbance by humans. Finally, published evaluations of the use of hunter observations to index moose populations have often reported that considering the size of a hunting group is necessary to improve the accuracy of those data; our analysis suggests an explanation – differences in escape behaviour between the sexes. ALCES VOL. 40: 123-131 (2004) Key words: escape behaviour, gender differences, group size, hunters, hunting, moose, move- ments, predator, reaction distance, selective force, Sweden Predation has long been considered one of the most important selective pressures on animals in the wild (Treves 2000). It is thus not surprising that ecologists have long studied many different aspects of predator and prey behaviour. For example, the ob- servation that the vigilance of individual prey declines as their group size increases is one of the most frequently reported rela- tionships in the study of animal behaviour (Roberts 1996). Among prey species in general, various aspects of escape or anti-predator behav- iour have been found to vary with a great number of factors including group size (Hebblewhite and Pletscher 2002), distance to cover (White and Berger 2001), tem- perature (Fernandez-Juricic et al. 2002), degree of predation risk (Hamilton and Heithaus 2001), experience with predators (Jachner 2001), and the sex of the prey (Magurran and Nowak 1991). Not surpris- ingly, evolution often has led to prey that are sensitive to the costs and benefits of differ- ent anti-predator behaviours under various circumstances (Frid and Dill 2002). Studies of moose (Alces alces) also have revealed a variety of anti-predator behaviours, that include standing its ground against wolves (Canis lupus, Mech 1970), increasing group MOOSE ESCAPE BEHAVIOUR – BASKIN ET AL. ALCES VOL. 40, 2004 124 size when foraging further from cover (Molvar and Bowyer 1994), selecting un- predictable sites to give birth (Bowyer et al. 1999), grouping together if predators are present, and increasing vigilance when with active young or when further from protec- tive refugia (White and Berger 2001, White et al. 2001). Franzmann and Schwartz (1998) provide a further overview of anti- predator (and many other) behaviours in North American moose. However, few published studies have quantified the spatial aspects of escape behaviour of moose when confronted by potential wild predators or humans (but see Glushkov 1976 and Andersen et al. 1996). Hunting is the most important mortality factor of moose in Fennoscandia (e.g., 81- 91% of adult mortality; Ericsson and Wallin 2001) because large predators (brown bears, Urus arctos, and wolves, Canis lupus) are absent over large areas or occur only at very low densities (Swenson et al. 1994, Persson and Sand 1998, Ericsson et al. 2001). Anecdotal reports by hunters and others suggest that there may be individual variation in escape behaviour – some moose use different escape tactics than others. Furthermore, at least in Russia, moose have been reported to exhibit different escape behaviour in areas with low and high hunt- ing pressure (Glushkov 1976). More knowl- edge on this issue is needed because hu- mans might be acting as a strong selective agent in those areas that have high hunting pressure, and thereby may alter moose be- haviour in the long term. Perhaps most importantly, we need to understand escape behaviour if we are to understand and man- age the interaction between the human predator and its prey. Here, we report what is apparently the first quantitative investigation to consider: (1) does escape behaviour from humans differ between the sexes in moose; (2) is there any correlation between age and the way a given moose attempts to escape; and (3) does group size affect escape behaviour in moose? METHODS The study area (~4,000 km2) is just north of Umeå (63°48’ N, 20°17’E), in coastal northern Sweden (Fig. 1). Norway spruce (Picea abies), Scots pine (Pinus sylvestris), and birches (Betula pendula and B. pubescens) are the dominant tree species, and the field layer vegetation is dominated by bilberry (Vaccinium myrtillis), lingonberry (V. vitis-idea), and heather (Calluna vulgaris). The length of the growing season is about 150 days, starting around mid-May; snow normally arrives in early November and persists to the end of April (SNA 1995). Moose density in the study area (as determined by helicopter census and pellet group counts) ranged between 0.7 and 0.9 moose km-2 (J.P. Ball unpublished data). Predators (other than humans) capable of taking moose essen- Fig. 1. Map of the study area in northern Sweden where moose were experimentally disturbed to quantitatively investigate the effects of age, sex, and group size on escape behaviour. Umeå is at the southern edge of the study area. ALCES VOL. 40, 2004 BASKIN ET AL. - MOOSE ESCAPE BEHAVIOUR 125 tially are absent (Swenson et al. 1994, Persson and Sand 1998). Moose are hunted heavily in the area and every year about one-third of the population is harvested. More importantly, in this area, hunters ac- count for 81% of all deaths of female moose born and 91% of all male deaths (Ball et al. 1999, Ericsson and Wallin 2001). Using a dart rifle fired from a helicop- ter, we immobilized adult and calf moose with a mixture of ethorphine and xylazine hydrochloride (Sandegren et al. 1987) dur- ing February to mid-March every year from 1990. Each moose was fitted with a num- bered radio collar and small numbered ear tags. Moose were aged using two comple- mentary methods: at marking, we deter- mined their ages using tooth wear and erup- tion (Skuncke 1949), and if later harvested, we obtained the jaw, sectioned the first molar, and counted the cementum annuli under at least 20x magnification (Sergeant and Pimlott 1959, Fancy 1980, Bubenik 1998). When good conditions for snow track- ing occurred between January and April during the 1996-97 and 1997-98 winters, we conducted controlled disturbance experi- ments with these known-aged radio-col- lared moose. For all tests, a single observer (for consistency) first located the moose by radio-telemetry and thereafter the moose (one individual or group of moose) was disturbed in a uniform way (i.e., a slow quiet approach on skis at the same speed from an initial distance of 70-100 m). After the disturbance, the observer followed the move- ments of the radio-collared moose by radio signals until it settled down (i.e., started to feed, ruminate, lie down, etc.). Then the observer tracked the moose backward in the snow to quantify the escape path taken (and avoid influencing the escape behaviour other than the initial controlled disturbance). We quantified the group size, and 5 response variables which quantified the spatial aspects of the moose’s escape be- haviour (for a group of moose, the centroid of the group was used). These were: (1) the distance the moose ran, trotted, or galloped after the initial disturbance (“Gallop Dis- tance” hereafter); (2) the total distance the moose moved before settling down (“Path Length to Quieting”); (3) the straight line distance from the place where the moose was disturbed to the point where the moose quieted down (“Straight Length to Quiet- ing”); (4) the number of stops the moose made before quieting (“Number of Stops Before Quieting”); and (5) we calculated a simple index of tortuosity (“Tortuosity”) by dividing the “Straight Length to Quieting” by the “Path Length to Quieting”. Thus, an animal with a perfectly linear escape path would have a Tortuosity Index of 1.0, whereas a moose which travelled in a non- linear fashion after being disturbed would have an index with a much lower value (e.g., 0.3). See Mårell et al. (2002) for additional approaches to quantifying move- ment. Indicator Variable Regression (also known as “Dummy Variable Regression”; Kleinbaum et al. 1987, Tabachnick and Fi- del 2001) was used to test for the effects of the independent variables (sex, group size, and age) on the dependent variables that described escape behaviour. Indicator Variable Regression is a form of a General Linear Model that is appropriate to a mix- ture of categorical (e.g., sex) and continu- ous (e.g., tortuosity) variables (Kleinbaum et al. 1987, Tabachnick and Fidel 2001). The significance factor reported is the addi- tional explanatory value of adding a variable to a model already containing the other two independent variables (i.e., this statistically controls for any correlations among these independent variables; Cohen and Cohen 1983, Kleinbaum et al. 1987, Tabachnick and Fidel 2001). In keeping with standard practices, if and only if the ANOVA for the MOOSE ESCAPE BEHAVIOUR – BASKIN ET AL. ALCES VOL. 40, 2004 126 overall indicator variable regression (i.e., containing all independent variables) was significant, did we then examine which indi- vidual effects (e.g., sex, group size, or age) accounted for this. Statistical analyses were performed with JMP version 4.0.5 (SAS Institute 2000). We report means ± 1 standard deviation unless otherwise noted. RESULTS In total, 29 controlled disturbances were performed (13 males and 16 females, in- cluding one cow of unknown age). The average age of the males we studied was 7.5 ± 2.2 years, and that of females 10.9 ± 4.5 years. The heavy hunting pressure in the study area is perhaps reflected by the rather strong escape responses made by moose to our experimental disturbances. The Gallop Distance was 293 ± 222 m, the Path Length to Quieting was 1,324 ± 699 m, and the Straight Length to Quieting was 855 ± 419m. Indicator Variable Regression revealed that the distance that moose ran after being disturbed (Gallop Distance) showed no re- lationship with the independent variables sex, age, or group size (whole model r2 = 0.08, P = 0.59), so we did not examine separate effects. In contrast, the overall model with the dependent variable Path Length to Quieting was significant (r2 = 0.36, P = 0.01); Table 1 reveals that age was not important, but sex and group size were predictors (larger groups travelled shorter distances before quieting) and fe- males travelled less (1,022m) before quiet- ing down than did males (1,671m). The overall model testing the explana- tory effects of sex, group size, and age on Straight Length to Quieting was not signifi- cant (whole model r2 = 0.23, P = 0.11), so again we do not consider separate effects. The overall model testing the effects of sex, age, and group size on the Number of Stops Before Quieting was significant (r2 = 0.33, P = 0.02). Here the independent variable that was important was group size (Table 1), with larger groups making fewer stops. Finally, the model testing the effects of sex, age, and group size on Tortuosity ex- plained nearly half of the observed variation (r2 = 0.46, P = 0.002). Here, however, group size and age were not predictors, but sex was a predictor (Table 1). Females exhibited more linear escape paths (mean Tortuosity Index = 0.78) than did males (mean Tortuosity Index = 0.57). Our analysis did not indicate an effect of age on escape behaviour. In contrast, our analysis revealed that group size was correlated with the Path Length to Quieting (larger groups settled down after travelling a shorter path). Furthermore, as group size increased, the group tended to keep moving (i.e., they made fewer stops). Finally, sev- eral important aspects of escape behaviour differed between the sexes. Although when initially disturbed both sexes galloped the same distance (Gallop Distance), and stopped the same linear distance away from Source df Sum of Squares F Ratio Prob > F sex 1 2030616.9 5.7058 0.0255 group size 1 1962848.5 5.5154 0.0278 age 1 139210.9 0.3912 0.5378 sex 1 18.336596 1.6672 0.2095 group size 1 55.928378 5.0851 0.034 age 1 42.879506 3.8986 0.0605 sex 1 0.40314635 19.282 0.0002 group size 1 0.00000142 0.0001 0.9935 age 1 0.04315248 2.0639 0.1643 Tortuosity Number of Stops Before Quieting Path Length to Quieting (m) Table 1. ANOVA effect tables from the indicator variable regression on moose (Alces alces) escape behaviour in Sweden. The results of the overall ANOVAs that were performed first are given in the text. Significant effects are highlighted in bold. ALCES VOL. 40, 2004 BASKIN ET AL. - MOOSE ESCAPE BEHAVIOUR 127 the initial disturbance point (Straight Length to Quieting), males travelled a greater total distance (1,671 ± 227 m) than females (1,022 ± 104 m) before settling down (Path Length to Quieting), and males did this by moving in a much more tortuous path than did females (Tortuosity Index for males 0.58 ± 0.06, females 0.78 ± 0.03). DISCUSSION Ungulates use a variety of different strategies to escape hunters. Hiding is a special behaviour exhibited by some forest- dwelling ungulates like wild boar and moose under extremely high hunting pressure (Baskin 1976, 1998). Moose also some- times escape by running to a safe distance and, from there, observe human activities (Glushkov 1976). This type of behaviour has also been reported for musk deer Moschus moschiferus (Zaitsev 1983), roe deer Capreolus capreolus (Danilkin 1996), reindeer Rangifer tarandus (Baskin and Hjältén 2001), and Alpine ibex Capra ibex (Krämer and Aeschbacher 1971). Often, animals demonstrate apparently reckless flight when they barely have seen or heard an approaching human. This type of flight behaviour is common under high hunting pressure for reindeer (Baskin and Hjältén 2001) and moose (Glushkov 2002). A more extreme escape behaviour is for the animal to run beyond its home range; e.g., wild camels Camelus bactrianus in Mongolia (Przewalsky 1878), European bison Bison bonasus in Caucasus at the beginning of the 20th century (Filatov 1912), and Ameri- can bison Bison bison in North America at the end of the 19th century (Roe 1951). Our finding that the Path Length to Quieting was 1,671m for males and 1,022m for females is comparable to the overall mean distance of 1,147m travelled by moose fleeing skiers and pedestrians in another rather heavily hunted Scandinavian moose population (Andersen et al. 1996). We were unable to find any published study on a lightly hunted population for comparison. Our analysis, that revealed the ten- dency of males to escape by a more tortu- ous path than females, may have implica- tions for their likelihood of escaping hunt- ers, but also for how moose counts should be performed, and of understanding the potential biases in hunter observations of moose. We suggest that the greater tortu- osity of the escape path of males may mean that they are more likely to come into re- peated contact with hunters, and thus more likely to be killed. However, compared to females, males are preferentially harvested by Swedish hunters anyway (Ball et al. 1999), so we are unable to separate the magnitude of the effects of these two fac- tors. In northern Sweden, the most com- mon method of hunting is through the use of baying dogs, whereby the moose stands its ground while it is distracted by a barking dog (and is subsequently shot by a hunter sneaking in) or it flees (and hunters sta- tioned around a particular parcel of land may be able to shoot the moose; Ball et al. 1999). Thus, hunting with dogs should provide a selective pressure for moose that flee, rather than hide, to escape harvest. In Sweden, hunting grounds are typically pri- vate, with a group of hunters having the right to hunt only on their particular hunting grounds. As a result, moose practicing a strategy of fleeing may escape the chain of shooters and find themselves beyond har- vest by a particular group of hunters. Be- cause moose hunting seasons are long (Sep- tember 1 – December 31), escaping moose often will find themselves in a new area where there is no group of hunters active on the same day. Compared to escaping in a linear fashion, remaining in the same area (and moving in a tortuous path) seems to be a poor strategy because of the scenting ability of the hunter’s dogs; nevertheless, our analysis reveals that this is how males MOOSE ESCAPE BEHAVIOUR – BASKIN ET AL. ALCES VOL. 40, 2004 128 We suggest moose escape behaviour is also influenced by group cohesiveness: when any group member (which our analysis sug- gests would often be a female) flees from a predator, so does the entire group. We suggest that moose managers should be aware of possible interactions among group size, hunting party size, and apparent sex ratio if using hunter observations to esti- mate moose population characteristics. Fur- ther research in this area should help clarify these relationships, and we suggest that incorporating escape behaviour of moose in a quantitative manner, as we have done, may be necessary to understand the mecha- nisms affecting population estimates. Our analysis provided no evidence of any relationship between age and escape behaviour, but our test of this aspect is not powerful because of sample size limita- tions. We suggest that future studies on escape behaviours focus on the effects of group size and sex because the age of wild moose cannot be judged from a distance and few studies will have the luxury, as we did, of tranquilizing moose first to determine their age. Having age data is ideal, but our analysis suggests that it is not essential to understand the escape behaviour of moose towards humans. The behavioural ecology literature sug- gests two main advantages for grouping in the presence of predators. First, vigilance: many eyes to detect predators (see Williams et al. 2003 and Bednekoff and Lima 1998 for recent overviews). Second, the other major advantage of being in a group comes from risk dilution: the risk to a given animal in a group of 2 is 50%, in a group of 100 animals it is 1% (Hamilton 1971, Treves 2000, Carbone et al. 2003). In accordance with the risk dilution hypothesis, our analy- sis reveals that moose apparently do feel safer as group size increases, as indicated by shorter travel distances to quieting. In conclusion, the differential mortality behave (perhaps the result of past selection by native predators). We suggest that, compared to females, males either prefer to keep a pursuing human within their sensory range or males have a stronger tendency to stay within their home range. Our analysis suggests that if the obser- vations of hunters are used to estimate population parameters, the proportion of males will be overestimated relative to their true fraction in the population (see Ericsson and Wallin 2001, Sylvén 2003 for additional analyses). Furthermore, if the differences between the sexes that we detected are indeed general, these behavioural differ- ences may help to explain why the size of a hunting group often emerges as an impor- tant factor in studies that use observations of moose seen per hour and per hunter, to estimate moose numbers (Ericsson and Wallin 2001, Sylvén 2003). We suggest that this may be because relative to smaller groups, larger groups of hunters may over- estimate males even more because the same bull may be sighted several times as it circles to remain in the area. We do not suggest that our controlled disturbance by a skier just after the hunting season precisely reflects the responses of moose to hunters during the hunt. Rather, our aim was to test if age, sex, or group size in moose is associated with differences in how moose attempt to escape from humans, and to do so in an area where humans are the main mortality source (and thus likely are acting as a selective force on moose behaviour). Thus, quantitative variables like the Distance to Quieting, for example, may well differ between the hunting season and the immediate post-hunting period. On the other hand, the qualitative relationships revealed by our analysis (e.g., the differen- tial responses between the sexes and the effect of group size) seem likely to be less temporally variable, although further re- search is needed to test this proposition. ALCES VOL. 40, 2004 BASKIN ET AL. - MOOSE ESCAPE BEHAVIOUR 129 exerted by hunters may be selecting for moose which employ a strategy of immedi- ate linear flight, or at least run out of sen- sory contact with a human after distur- bance. We suggest that escape behaviour is an under-studied aspect of moose ecol- ogy. This is surprising since humans ac- count for over 80% of all moose deaths in Sweden (Ericsson and Wallin 2001), and are thus expected to be a rather strong selective force. Furthermore, understand- ing escape behaviour may be critical in improving our ability to use hunter observa- tions to estimate moose populations. Spe- cifically, differences between the sexes in escape behaviour may lead to an apparently greater proportion of males being detected in hunter observations. The variation in escape behaviour with group size we de- tected may explain why the size of a hunting party often emerges as a significant factor in previous analyses that use hunter obser- vations (e.g., Ericsson and Wallin 2001, Sylvén 2003). Based on our analysis, we suggest that quantifying the escape behav- iour of moose may lead to a better under- standing of the selective force human hunt- ers are exerting on moose populations, and may lead to improvements in using hunter observations to manage moose populations. ACKNOWLEDGEMENTS We thank the Swedish University of Agricultural Sciences (grant to Kjell Danell) for financial support for Leonid Baskin dur- ing his stay in Umeå, and the Swedish Environment Protection Agency (grants to Kjell Danell and John P. Ball) for financial support. We also gratefully acknowledge the tireless and expert work by our field technicians (particularly Åke Nordström, E r i c A n d e r s s o n , a n d N i l s - G u n n a r Andersson) in the moose project, and K. Wallin for his early work in building up the marked moose population for related stud- ies. The enthusiastic co-operation of many hundreds of moose hunters over more than a decade allowed us to exactly age our moose, and made possible this and many other investigations. We thank T.R. Stephenson, R. Rea, and an anonymous reviewer for comments which greatly im- proved the manuscript. REFERENCES ANDERSEN, R., J. D. C. LINELL, and R. LANGVATN. 1996. Short term behav- ioural and physiological response of moose Alces alces to military distur- bance in Norway. Biological Conser- vation 77:169-176. BALL, J. P., G. ERICSSON, and K. WALLIN. 1999. Climate change, moose and their human predators. Ecological Bulletins 47:178-187. BASKIN, L. M. 1976. Behaviour of Ungu- lates. Nauka publishers, Nauka, Mos- cow, Russia. _____. 1998. Moose conservation in eco- systems of Eastern Europe. Alces 34:395-407. _____, and J. HJÄLTEN. 2001. Fright and flight behaviour of reindeer. Alces 37:435-446. BEDNEKOFF, P. A., and S. L. LIMA. 1998. Re-examining safety in numbers: inter- actions between risk dilution and col- lective detection depend upon predator targeting behaviour. Proceedings of the Royal Society of London Series B- Biological Sciences 265:2021-2026. BOWYER R. T., V. VAN BALLENBERGHE, J. G. KIE, and J. A. K. MAIER. 1999. Birth- site selection by Alaskan moose: mater- nal strategies for coping with a risky environment. Journal of Mammalogy 80:1070-1083. BUBENIK, A. B. 1998. Evolution, taxonomy and morphophysiology. Pages 77-124 in A. W. Franzmann and C.C. Schwartz, editors. Ecology and Management of t h e N o r t h A m e r i c a n M o o s e . MOOSE ESCAPE BEHAVIOUR – BASKIN ET AL. ALCES VOL. 40, 2004 130 Smithsonian Institution Press, Wash- ington, D.C., USA. CARBONE C., W. A. THOMPSON, I. ZADORINA, and J. M. Rowcliffe. 2003. Competi- tion, predation risk and patterns of flock expansion in barnacle geese (Branta leucopsis). Journal of Zoology 259:301- 308. COHEN, J., and P. COHEN. 1983. Applied Multiple Regression Correlation Analy- s i s f o r t h e B e h a v i o r a l S c i e n c e s . Erlbaum, New York, New York, USA. DANILKIN, A. A. 1996. Behavioral Ecology of Siberian and European Roe Deer. Chapman and Hall, London, U.K. ERICSSON, G., and K. WALLIN. 2001. Age- specific moose (Alces alces) mortality in a predator-free environment: evidence for senescence in females. Ecoscience 8:157-163. _____, K. WALLIN, J. P. BALL, and M. BROBERG. 2001. Age-related reproduc- tive effort and senescence in free-rang- ing moose Alces alces. Ecology 82:1613-1620. FANCY, S. G. 1980. Preparation of mamma- lian teeth for age determination by ce- mentum layers: a review. Wildlife So- ciety Bulletin 8:242-248. FERNANDEZ-JURICIC, E., M. D. JIMENEZ, and E. LUCAS. 2002. Factors affecting intra- and inter-specific variations in the difference between alert distances and flight distances for birds in forested habitats. Canadian Journal of Zoology 80:1212-1220. FILATOV, D. 1912. Caucasian bison. Zapiski imperatorskoi akademii nauk, otdelenie phisiko-matematicheskoe, Series 7, 30:1- 40. (In Russian). FRANZMANN, A. W., and C. C. SCHWARTZ, editors. 1998. Ecology and Manage- ment of the North American Moose. Smithsonian Institution Press, Wash- ington, D.C., USA. FRID, A., and L. DILL. 2002. Human- caused disturbance stimuli as a form of predation risk. Conservation Ecology 6:11-16. GLUSHKOV, V. M. 1976. Some peculiarities of moose behavior that determine hunter’s success. In: S.A. Korytin, editor. Behavior of Hunting Animals. S b o r n i k n a u c h n o - t e k h n i c h e s k o y informatsii 51-52:51-58. (In Russian). _____. 2002. Ecological basis of moose population management in Russia. The- sis. Vsesoyuznyi Naychnoissledovatel- sky institut okhotnichego khozyaistva i zverovodstva, Kirov, Russia. (In Rus- sian). HAMILTON, I. M., and M. R. HEITHAUS. 2001. The effects of temporal variation in predation risk on anti-predator behav- iour: an empirical test using marine snails. Proceedings of the Royal Society of London Series B-Biological Sciences 268:2585-2588. HAMILTON, W. D. 1971. Geometry of the selfish herd. Journal of Theoretical Biology 31:295-311. HEBBLEWHITE, M., and D. PLETSCHER. 2002. Effects of elk group size on predation by wolves. Canadian Journal of Zool- ogy 80:800-809. JACHNER, A. 2001. Anti-predator behaviour of naive compared with experienced juvenile roach. Journal of Fish Biology 59:1313-1322. KLEINBAUM, D. G., L. L. KUPPER, and K. E. MULLER. 1987. Applied Regression Analysis and Other Multivariate Meth- ods. Duxbury Press, Belmont, Califor- nia, USA. KRÄMER, A., and A. AESCHBACHER. 1971. Zum Slachterhalten des Steinwildes (Capra ibex) im Oberengadin, Schweiz. Säugetierk Mitteilungen 19:164-171. MAGURRAN, A. E., and M. A. NOWAK. 1991. Another battle of the sexes - The con- sequences of sexual asymmetry in mat- ing costs and predation risk in the guppy, ALCES VOL. 40, 2004 BASKIN ET AL. - MOOSE ESCAPE BEHAVIOUR 131 Poecilia reticulata. Proceedings of the Royal Society of London Series B- Biological Sciences 246:31-38. MÅRELL, A., J. P. BALL, and A. HOFGAARD. 2002. Foraging and movement paths of female reindeer: insights from fractal analysis, correlated random walks and Lévy flights. Canadian Journal of Zool- ogy 80:854-865. MECH, L. D. 1970. The Wolf: The Ecology and Behavior of an Endangered Spe- cies. Natural History Press, Garden City, New York, USA. MOLVAR, E. M., and R. T. BOWYER. 1994. Costs and benefits of group living in a recently social ungulate - the Alaskan moose. Journal of Mammalogy 75:621- 630. PERSSON, J., and H. SAND. 1998. Vargen — viltet, ekologin och människan. Svenska Jägareforbundet, Spånga, Sweden. (In Swedish). PRZEWALSKY, N. M. 1878. Wild camel. Priroda I okhota, 1-2. ROBERTS, G. 1996. Why individual vigilance behaviour declines as group size in- creases. Animal Behaviour 51:1077- 1086. ROE, F. G. 1951. The North American Buffalo. A Critical Study of the Spe- cies in its Wild State. University of Toronto Press, Toronto, Ontario, Canada. SANDEGREN, F., L. PETTERSSON, P. AHLQVIST, and B. O. RÖKEN, 1987. Immobilization of moose in Sweden. Swedish Wildlife Research Supplement 1:785-791. SAS Institute. 2000. JMP Statistics and Graphics Guide. SAS Institute, Cary, North Carolina, USA. SERGEANT, D. E., and D. H. PIMLOTT. 1959. Age determination in moose from sec- tioned incisor teeth. Journal of Wildlife Management 23:315-321. SKUNCKE, F. 1949. Älgen, studier, jakt och vård. P. A. Nordstedts och Söners Förlag, Stockholm, Sweden. (In Swed- ish). (SNA) Sveriges National Atlas. 1995. Sveriges National Atlas - Klimat, sjöar och vattendrag. Bra Böcker, Höganäs, Sweden. SWENSON, J. E., F. SANDEGREN, A. BJÄRVALL, A. SÖDERBERG, P. WABAKKEN, and R. FRANZÉN. 1994. Size, trend, distribution and conservation of the brown bear Ursus arctos population in Sweden. Biological Conservation 70:9-17. SYLVÉN, S. 2003. Management and regu- lated harvest of moose (Alces alces) in Sweden. Agraria 371. Doctoral The- sis, Department of Conservation Biol- ogy, Swedish University of Agricultural Sciences, Uppsala, Sweden. TABACHNICK, B. G., and L. S. FIDELL. 2001. Using Multivariate Statistics. Fourth edition. Allyn and Bacon, Needham Heights, Massachusetts, USA. TREVES, A. 2000. Theory and method in studies of vigilance and aggregation. Animal Behaviour 60:711-722. WH I T E , K. S., and J. BE R G E R. 2001. Antipredator strategies of Alaskan moose: are maternal trade-offs influ- enced by offspring activity? Canadian Journal of Zoology 79:2055-2062. _____, J.W. TESTA, and J. BERGER. 2001. Behavioral and ecologic effects of dif- ferential predation pressure on moose in Alaska. Journal of Mammalogy 82:422-429. WILLIAMS, C. K., R. S. LUTZ, and R. D. APPLEGATE. 2003. Optimal group size and northern bobwhite coveys. Animal Behaviour 66:377-387. ZAITSEV, V. A. 1983. Escape behaviour of musk deer (Moschus moschiferus parviceps) in middle Sikhote-Alin. Zoologichesky Zhurnal 62:1718-1726.